6 numbers your competitor publishes by accident

No competitor is going to publish their ARR. But almost every one of them publishes a rate limit, a seat default, a member count, or a directory that can be counted. These are not metrics in the strategic sense. They are operational side effects — numbers that had to be true for the page to work.

That makes them harder to fake than a positioning statement. A marketing claim costs nothing to write. A rate limit costs engineering time to change.

Here are six worth counting, and what a change in each one actually means.

1. The default seat count in their pricing calculator

Open the pricing page, find the slider or the quantity field, and note where it starts. Somebody chose that number. It is almost always tuned to the deal size they want to look normal — high enough to make the plan price look reasonable, low enough not to scare a solo evaluator.

When the default moves from 5 to 20, the page is no longer written for the same buyer. That is a positioning change that never appears in a changelog and rarely appears in a press release. Watch it alongside the ninety-day view of pricing page diffs — the slider often moves a release or two before the tiers do.

2. The rate limits in their API docs

Rate limits are capacity statements. A vendor who raises the default from 100 requests per minute to 1,000 has either bought headroom or rewritten something expensive. A vendor who lowers one has found a cost problem.

The per-tier split matters more than the absolute number. If enterprise gets 50x the free tier, they are selling volume. If every tier gets the same limit and the difference is in features, they are selling capability. That distinction tells you which objection their sales team is trained to handle.

3. The member count on their community

Slack, Discord, and Circle all display a member number, and most vendors leave it visible because it flatters them. Sample it monthly and you have a growth curve nobody meant to give you.

The rate is the signal, not the total. Forty thousand members added over four years is a different company than forty thousand added in eight months. And when the curve flattens while their hiring accelerates, something in the funnel stopped working. Mining the community for pain tells you what customers are struggling with; counting it tells you how many of them there are.

4. "Trusted by N teams"

The homepage number. It is the softest of the six — definitions of "team" are elastic, and nobody audits it. But it is still a number a legal team signed off on, and the update cadence is informative even when the value is not.

A company that refreshes it quarterly is tracking it. A company whose "10,000+ teams" has been unchanged for two years has either stopped growing or stopped caring about the page — and you can tell which by checking whether anything else on the page moved. If the whole page is frozen, it is neglect. If everything moved except the number, it is a decision.

5. The integration directory count

Count the entries. Then count them again in a month. Integration directories are one of the few places where a company's roadmap is expressed in a number that goes up.

Direction of travel matters both ways. Additions cluster around a strategy — five CRM connectors in a quarter means they are chasing a segment. Removals are louder. A vendor who quietly drops eight integrations has either deprecated a platform or run out of the maintenance budget to keep them alive, and neither gets announced.

6. Changelog entries per month

Not what shipped — how often. Count the dated entries per month for the last year and you get a shipping cadence you can compare across competitors and against their own history.

A cadence that drops by half and stays there usually means engineering got reallocated: to a rewrite, to an enterprise deal, to a platform migration. A cadence that doubles right before a funding announcement is a different story. Either way, the count is more reliable than the content, because entries get padded but frequency is hard to fake.

Counting is the easy part

The tedious part is doing it on the same day every month, for every competitor, without skipping the quiet ones. That is where most manual tracking dies — not in the analysis, but in the discipline of the sample.

This is the shape of problem Seeto is built for: it watches public surfaces continuously and surfaces what changed as a discrete event, so a rate limit that moved in March is on the record in March rather than discovered in July. It will not tell you what a number means — that read is yours, and it needs the context of the market you actually sell into. But it will make sure you have both endpoints, which is the whole game when the value of a number is the fact that it moved. If you only have the current value, you don't have a signal; you have a screenshot, and a screenshot is not a change record.

Pick two of the six and start a spreadsheet this week. In three months you will have something no competitor briefing in your company currently contains: a trend line.